Number Systems and Binary Arithmetic

Quantitative Analysis II Professor Bob Orr

Introduction to Numbering Systems

We are all familiar with the decimal number system (Base 10). Some other number systems that we will work with are:
Binary Base 2 Octal Base 8 Hexadecimal Base 16

Copyright 2000 Indiana University Board of Trustees

Characteristics of Numbering Systems

1)
2) 3) 4) 5) 6)

The digits are consecutive. The number of digits is equal to the size of the base. Zero is always the first digit. The base number is never a digit. When 1 is added to the largest digit, a sum of zero and a carry of one results. Numeric values determined by the have implicit positional values of the digits.
Copyright 2000 Indiana University Board of Trustees

Significant Digits
Binary: 11101101
Most significant digit Least significant digit

Hexadecimal: 1D63A7A
Most significant digit Least significant digit

Copyright 2000 Indiana University Board of Trustees

Binary Number System

Also called the Base 2 system The binary number system is used to model the series of electrical signals computers use to represent information 0 represents the no voltage or an off state 1 represents the presence of voltage or an on state

Copyright 2000 Indiana University Board of Trustees

Binary Numbering Scale

Base 2 Number Base 10 Equivalent Power Positional Value

000 001 010 011 100 101 110 111

0 1 2 3 4 5 6 7

20 21 22 23 24 25 26 27

1 2 4 8 16 32 64 128

Copyright 2000 Indiana University Board of Trustees

Binary Addition
4 Possible Binary Addition Combinations:

(1)
Carry

(3)

0 +0 00 1 +0 01

(2)
Sum

(4)

0 +1 01 1 +1 10

Note that leading zeroes are frequently dropped.

Copyright 2000 Indiana University Board of Trustees

Decimal to Binary Conversion

The easiest way to convert a decimal number to its binary equivalent is to use the Division Algorithm This method repeatedly divides a decimal number by 2 and records the quotient and remainder
The remainder digits (a sequence of zeros and ones) form the binary equivalent in least significant to most significant digit sequence
Copyright 2000 Indiana University Board of Trustees

Division Algorithm
Convert 67 to its binary equivalent:
6710 = x2
Step 1: 67 / 2 = 33 R 1 Step 2: 33 / 2 = 16 R 1 Step 3: 16 / 2 = 8 R 0 Step 4: 8 / 2 = 4 R 0 Step 5: 4 / 2 = 2 R 0 Step 6: 2 / 2 = 1 R 0 Step 7: 1 / 2 = 0 R 1 Divide 67 by 2. Record quotient in next row Again divide by 2; record quotient in next row Repeat again Repeat again Repeat again Repeat again STOP when quotient equals 0

1 0 0 0 0 1 12
Copyright 2000 Indiana University Board of Trustees

Binary to Decimal Conversion

The easiest method for converting a binary number to its decimal equivalent is to use the Multiplication Algorithm Multiply the binary digits by increasing powers of two, starting from the right Then, to find the decimal number equivalent, sum those products

Copyright 2000 Indiana University Board of Trustees

Multiplication Algorithm
Convert (10101101)2 to its decimal equivalent:
Binary Positional Values Products

1 0 1 0 1 1 0 1

x x x x x x x x

27 26 25 24 23 22 21 20
128 + 32 + 8 + 4 + 1

17310
Copyright 2000 Indiana University Board of Trustees

Octal Number System

Also known as the Base 8 System Uses digits 0 - 7 Readily converts to binary Groups of three (binary) digits can be used to represent each octal digit Also uses multiplication and division algorithms for conversion to and from base 10

Copyright 2000 Indiana University Board of Trustees

Decimal to Octal Conversion

Convert 42710 to its octal equivalent:
427 / 8 = 53 R3 53 / 8 = 6 R5 6 / 8 = 0 R6 Divide by 8; R is LSD Divide Q by 8; R is next digit Repeat until Q = 0

6538
Copyright 2000 Indiana University Board of Trustees

Octal to Decimal Conversion

Convert 6538 to its decimal equivalent:
Octal Digits

6
x

5 81

Positional Values
Products

3 80

82

384 + 40 + 3

42710
Copyright 2000 Indiana University Board of Trustees

Octal to Binary Conversion

Each octal number converts to 3 binary digits
To convert 6538 to binary, just substitute code: 6 5 3

110 101 011

Copyright 2000 Indiana University Board of Trustees

Hexadecimal Number System

Base 16 system Uses digits 0-9 & letters A,B,C,D,E,F Groups of four bits represent each base 16 digit

Copyright 2000 Indiana University Board of Trustees

Decimal to Hexadecimal Conversion

Convert 83010 to its hexadecimal equivalent:
830 / 16 = 51 R14 51 / 16 = 3 R3 3 / 16 = 0 R3
= E in Hex

33E16
Copyright 2000 Indiana University Board of Trustees

Hexadecimal to Decimal Conversion

Convert 3B4F16 to its decimal equivalent:
Hex Digits

3
x

B
x

Positional Values
Products

F
x

163 162 161 160 12288 +2816 + 64 +15

15,18310
Copyright 2000 Indiana University Board of Trustees

Binary to Hexadecimal Conversion

The easiest method for converting binary to hexadecimal is to use a substitution code Each hex number converts to 4 binary digits

Copyright 2000 Indiana University Board of Trustees

Substitution Code
Convert 0101011010101110011010102 to hex using the 4-bit substitution code :
5 6 A E 6 A

0101 0110 1010 1110 0110 1010

56AE6A16
Copyright 2000 Indiana University Board of Trustees

Substitution Code
Substitution code can also be used to convert binary to octal by using 3-bit groupings:
2 5 5 2 7 1 5 2

010 101 101 010 111 001 101 010

255271528
Copyright 2000 Indiana University Board of Trustees

Complementary Arithmetic

1s complement
Switch all 0s to 1s and 1s to 0s

Binary # 1s complement

10110011 01001100

Copyright 2000 Indiana University Board of Trustees

Complementary Arithmetic

2s complement
Step 1: Find 1s complement of the number Binary # 11000110 1s complement 00111001 Step 2: Add 1 to the 1s complement 00111001 + 00000001 00111010
Copyright 2000 Indiana University Board of Trustees

Signed Magnitude Numbers

110010.. 00101110010101

Sign bit 0 = positive 1 = negative

31 bits for magnitude

This is your basic Integer format

Copyright 2000 Indiana University Board of Trustees

Floating Point Numbers

Real numbers must be normalized using scientific notation: 0.1 2n where n is an integer

Note that the whole number part is always 0 and the most significant digit of the fraction is a 1 ALWAYS!
Copyright 2000 Indiana University Board of Trustees

Floating Point Operations

Before two floating point numbers can be added, the exponents for both numbers must be made equal hence the term floating point NIST Standard Format (32-bit word)

8-bit exponent

23-bit fraction field

Copyright 2000 Indiana University Board of Trustees

Bias Notation
The exponent field (8 bits) can be used to represent integers from 0-255 Because of the need for negative exponents to be represented as well, the range is offset or biased from 128 to + 127 In this way, both very large and very small numbers can be represented

Copyright 2000 Indiana University Board of Trustees

Double Precision

Double word format that increases both the length of the fraction (precision) but also the size of the bias (magnitude)
NIST Standard format (64 bits)

10-bit exponent

53-bit fraction field

Copyright 2000 Indiana University Board of Trustees

Error Considerations

The Hole at Zero Problem

Regardless of how large an exponent field is, there are still smaller positive and negative numbers about zero that cannot be represented In bias notation, +2-129 and - 2-129 are beyond the acceptable range The Hole at Zero defines the range of values near zero that cannot be stored
Copyright 2000 Indiana University Board of Trustees

Computational Errors

When converting base 10 fractions to binary, only those fractions whose values can be expressed as a sum of base 2 fractions will convert evenly All other base 10 fractions feature a least significant bit that is either rounded or truncated an approximation When two such numbers are multiplied, the rounding error is compounded
Copyright 2000 Indiana University Board of Trustees